ClawHub
Back to skill

Security audit

Lattice

Security checks across malware telemetry and agentic risk

Overview

This appears to be a Lattice-focused skill with one scoping caveat, but no evidence of hidden execution, malware, persistence, or unrelated data access.

Before installing, be aware that Lattice may contain private employee and HR data. Use the skill for explicit Lattice tasks, and require confirmation before reading sensitive records, listing many users, exporting data, or making changes.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Vague Triggers

Medium
Confidence
93% confidence
Finding
The skill description and frontmatter instruct the agent to use this skill for ANY Lattice-related request, creating an overly broad routing rule without explicit task boundaries. This can cause the agent to invoke the connector for requests that may be out of scope, privacy-sensitive, or better handled with additional user confirmation, increasing the chance of unnecessary data access.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.